**3. Empirical results**

In the analysis we use two variables, an uncertainty index and a firm performance measure (MFA score). As these two variables are composite measures, although the methodology allow, we do not control other variables in the regressions. Based on Schwarz information criteria (SIC), we determine the number of lags included in the VAR model as two. In addition, following Rossi and Wang [39], we choose trimming parameter as 0.15 which is commonly used in structural break literature. After the estimation, we plot Wald test statistics through time which shows the strength of timevarying Granger causality from uncertainty measure to the firm performance measure.

We start our analysis by looking at the predictive ability of economic uncertainty on MFA score. Economic uncertainty index is constructed from five sub-uncertainty indices (forecasters' uncertainty index, financial uncertainty index, firms' uncertainty index, consumers' uncertainty index, and economic policy uncertainty index) via dynamic factor models explained by Cosar and Sahinoz [37].

**Table 4** shows the results of classical Granger causality test for VAR (2) model which indicates that there is a causality only at 10% level for the whole sample. Moreover, traditional Granger causality test suffers from instabilities in the data and does not show time variability in causality which is the critical point that this chapter tries to address.

**Figure 2** gives Wald statistics over time, which shows that the predictive power of economic uncertainty is increasing during 2008 global financial crisis. By choosing trimming parameter as 0.15, we are losing 16 observations from the beginning and end of the sample period. Hence, the latest observation is the 2016Q1 in this case, which leads us not to observe the recent financial distress period.

**Figure 2** depicts that there is a significant Granger causality from economic uncertainty towards balance sheet performance of BIST firms in our sample. The global financial crisis is associated with significant contraction in economic activity, rise in unemployment, and fall in investment and consumption spending in Turkish economy. Year-on-year growth rate of GDP dropped as high as (−5) percent in 2009. These adverse developments in that era caused firms' financial health to worsen, as could be monitored through declining MFA scores. Following global crisis, Turkish economy did not experience substantial shrinkage in the economic activity till 2016, which would damage the firms' performance. Some temporary shocks such as European debt crisis or tapering tantrum have been alleviated by the effective fiscal and monetary policies. The statistically insignificant values of Wald


**171**

**Figure 3.**

*SIC, trimming 3%.*

**Figure 2.**

*SIC, trimming 15%.*

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength*

statistics after the global crisis indicate that, in normal times, there is no causality

*Wald statistics of the test, H0: economic uncertainty does not Granger-cause MFA score. Notes: VAR(2) under* 

In order to check the effect of Turkey-specific currency volatility observed in 2018 on firms' scores, we take trimming parameter as 0.03, which allows us not to lose any observation. With the coverage of data for the year 2018, **Figure 3** indicates that there is a jump in Wald statistics meaning that Granger causality occurs at this period as well. This local-sourced turbulent period in 2018 led Turkish lira to devaluate by nearly 35 percent against US dollar. Due to the rising volatility in exchange rate, inflation soared, interest rates climbed up, and economic activity

Turkish firms accumulated large foreign currency debt after the global crisis since the expansionary monetary policies in advanced countries caused Turkish banks to reach foreign currency funding at low costs and with long maturities.

*Wald statistics of the test, H0: economic uncertainty does not Granger-cause MFA score. Notes: VAR(2) under* 

from economic uncertainty to firm performances.

contracted, which led to a mounting economic policy uncertainty.

*DOI: http://dx.doi.org/10.5772/intechopen.91860*

**Table 4.**

*Granger causality test for the whole sample.*

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength DOI: http://dx.doi.org/10.5772/intechopen.91860*

#### **Figure 2.**

*Banking and Finance*

the null hypothesis in Eq. (5):

**3. Empirical results**

chapter tries to address.

*yt* = β1,*<sup>t</sup> yt*−1 + β2,*<sup>t</sup> yt*−2 + …+ β*m*,*<sup>t</sup> yt*<sup>−</sup>*<sup>m</sup>* + ε*<sup>t</sup>* (4)

*H*0: β*<sup>t</sup>* = 0,*for all t* = 1,…,*T* (5)

where β*j*,*t*,*j* = 1,…*m* show time-varying coefficients, [*yt*, *yt*−1,…, *yt*<sup>−</sup>*m*]′ show variables in the model, and ε*t* shows heteroskedastic and serially correlated error term. Finally, based on Eq. (4), robust Granger causality test figures out the validity of

In the analysis we use two variables, an uncertainty index and a firm performance measure (MFA score). As these two variables are composite measures, although the methodology allow, we do not control other variables in the regressions. Based on Schwarz information criteria (SIC), we determine the number of lags included in the VAR model as two. In addition, following Rossi and Wang [39], we choose trimming parameter as 0.15 which is commonly used in structural break literature. After the estimation, we plot Wald test statistics through time which shows the strength of timevarying Granger causality from uncertainty measure to the firm performance measure. We start our analysis by looking at the predictive ability of economic uncertainty

on MFA score. Economic uncertainty index is constructed from five sub-uncertainty indices (forecasters' uncertainty index, financial uncertainty index, firms' uncertainty index, consumers' uncertainty index, and economic policy uncertainty

**Table 4** shows the results of classical Granger causality test for VAR (2) model which indicates that there is a causality only at 10% level for the whole sample. Moreover, traditional Granger causality test suffers from instabilities in the data and does not show time variability in causality which is the critical point that this

**Figure 2** gives Wald statistics over time, which shows that the predictive power of economic uncertainty is increasing during 2008 global financial crisis. By choosing trimming parameter as 0.15, we are losing 16 observations from the beginning and end of the sample period. Hence, the latest observation is the 2016Q1 in this

**Figure 2** depicts that there is a significant Granger causality from economic uncertainty towards balance sheet performance of BIST firms in our sample. The global financial crisis is associated with significant contraction in economic activity, rise in unemployment, and fall in investment and consumption spending in Turkish economy. Year-on-year growth rate of GDP dropped as high as (−5) percent in 2009. These adverse developments in that era caused firms' financial health to worsen, as could be monitored through declining MFA scores. Following global crisis, Turkish economy did not experience substantial shrinkage in the economic activity till 2016, which would damage the firms' performance. Some temporary shocks such as European debt crisis or tapering tantrum have been alleviated by the effective fiscal and monetary policies. The statistically insignificant values of Wald

**Zero hypothesis (H0) Chi2 P-value Decision** Econ uncertainty does not Granger-cause MFA score 5.8797 0.053 Do not reject

index) via dynamic factor models explained by Cosar and Sahinoz [37].

case, which leads us not to observe the recent financial distress period.

**170**

**Table 4.**

*Granger causality test for the whole sample.*

*Wald statistics of the test, H0: economic uncertainty does not Granger-cause MFA score. Notes: VAR(2) under SIC, trimming 15%.*

statistics after the global crisis indicate that, in normal times, there is no causality from economic uncertainty to firm performances.

In order to check the effect of Turkey-specific currency volatility observed in 2018 on firms' scores, we take trimming parameter as 0.03, which allows us not to lose any observation. With the coverage of data for the year 2018, **Figure 3** indicates that there is a jump in Wald statistics meaning that Granger causality occurs at this period as well. This local-sourced turbulent period in 2018 led Turkish lira to devaluate by nearly 35 percent against US dollar. Due to the rising volatility in exchange rate, inflation soared, interest rates climbed up, and economic activity contracted, which led to a mounting economic policy uncertainty.

Turkish firms accumulated large foreign currency debt after the global crisis since the expansionary monetary policies in advanced countries caused Turkish banks to reach foreign currency funding at low costs and with long maturities.

#### **Figure 3.**

*Wald statistics of the test, H0: economic uncertainty does not Granger-cause MFA score. Notes: VAR(2) under SIC, trimming 3%.*

Banks utilized these funds as loans to real sector firms, even the ones without any FX income. Firms without FX income lived with considerable FX mismatches in their books. As a result, the sharp depreciation of Turkish lira in 2018 created a substantial deterioration in the balance sheets of real sector, and many firms defaulted since they could not repay their FX debts. In that period, MFA scores have dropped as high as the fall in global crisis.

The rising Wald statistics in that period is an evidence of the significant causality from economic uncertainty to MFA scores as observed during the period of global crisis. As **Figure 3** suggests, there is no significant causality in normal times. Our overall evaluation from **Figures 2** and **3** is that when economic uncertainty accelerates, it significantly Granger-causes MFA scores to drop; nevertheless, when uncertainty is stable, there is no significant causality to firm performance.

One may question whether this causal relationship from economic uncertainty to firm performance can differentiate among different firm types. To answer this, we split the firms in our sample as exporters and domestic producers and check which type of uncertainty is important for exporters and non-exporters. Firstly, we check the causality from financial uncertainty to MFA scores. Financial uncertainty is a measure of volatility in global financial conditions such as VIX, EMBI, CDS, and exchange rate uncertainty [5]. One would expect that exporters are affected more from global financial conditions measured by the financial uncertainty index. Although exchange rate volatility is crucial for all economic agents, its effect is even higher for exporters. **Figures 4** and **5** show the results for exporters and nonexporters, respectively. While financial uncertainty has a strong effect in exporters' score in 2008 and 2018 volatility periods, none of those periods have a significant effect on domestic producers.

The potential uncertainty that can affect domestic producers is the uncertainty measuring the perception of consumers in Turkish economy. Hence, we estimate VAR(2) model with consumer uncertainty for exporters and non-exporters (**Figures 6–8**). Because of the big jump in 2005 in **Figure 6**, the explanatory power of consumer uncertainty on exporters is not clear with a trimming parameter of 3 percent. By using a trimming parameter of 0.15, we eliminate this period and see the exact relationship in **Figure 7**. As it is clear in the figure, the relationship is far from 10% significance level in the whole sample period for exporter firms. This

#### **Figure 4.**

*Wald statistics values of the test, H0: financial uncertainty does not Granger-cause MFA score for exporter. Notes: VAR(2) under SIC, trimming 3%.*

**173**

firms' performance.

*Notes: VAR(2) under SIC, trimming 15%.*

**Figure 6.**

**Figure 5.**

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength*

might be an expected result since perception of domestic consumers should not have an impact on exporting firms selling abroad. Global financial conditions or global consumer uncertainties are anticipated to have a more impact on exporter

*Wald Statistics Values of the Test, H0: consumer uncertainty does not Granger-cause MFA score for exporters.* 

*Wald statistics values of the test, H0: financial uncertainty does not Granger-cause MFA score for non-*

On the other hand, as seen in **Figure 8**, Wald statistics improved extensively after 2009 and exceeded 10% critical levels at some points till 2015 for non-exporter firms. The period between 2010 and 2015 is associated with a sound economic growth with a low-level of interest rates and stable exchange rate. In that period, households' perceptions and future expectations were robust, while their wealth was also improving. This led to soaring consumption expenditures, which increased the sales of non-exporter firms. As a result, the balance sheet performance of nonexporter firms significantly strengthened. We infer from those developments that, as consumers' perception worsens, this does not lead to a causality from consumer

*DOI: http://dx.doi.org/10.5772/intechopen.91860*

*exporters. Notes: VAR(2) under SIC, trimming 3%.*

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength DOI: http://dx.doi.org/10.5772/intechopen.91860*

#### **Figure 5.**

*Banking and Finance*

as high as the fall in global crisis.

effect on domestic producers.

Banks utilized these funds as loans to real sector firms, even the ones without any FX income. Firms without FX income lived with considerable FX mismatches in their books. As a result, the sharp depreciation of Turkish lira in 2018 created a substantial deterioration in the balance sheets of real sector, and many firms defaulted since they could not repay their FX debts. In that period, MFA scores have dropped

The rising Wald statistics in that period is an evidence of the significant causality from economic uncertainty to MFA scores as observed during the period of global crisis. As **Figure 3** suggests, there is no significant causality in normal times. Our overall evaluation from **Figures 2** and **3** is that when economic uncertainty accelerates, it significantly Granger-causes MFA scores to drop; nevertheless, when

One may question whether this causal relationship from economic uncertainty to firm performance can differentiate among different firm types. To answer this, we split the firms in our sample as exporters and domestic producers and check which type of uncertainty is important for exporters and non-exporters. Firstly, we check the causality from financial uncertainty to MFA scores. Financial uncertainty is a measure of volatility in global financial conditions such as VIX, EMBI, CDS, and exchange rate uncertainty [5]. One would expect that exporters are affected more from global financial conditions measured by the financial uncertainty index. Although exchange rate volatility is crucial for all economic agents, its effect is even higher for exporters. **Figures 4** and **5** show the results for exporters and nonexporters, respectively. While financial uncertainty has a strong effect in exporters' score in 2008 and 2018 volatility periods, none of those periods have a significant

The potential uncertainty that can affect domestic producers is the uncertainty measuring the perception of consumers in Turkish economy. Hence, we estimate VAR(2) model with consumer uncertainty for exporters and non-exporters (**Figures 6–8**). Because of the big jump in 2005 in **Figure 6**, the explanatory power of consumer uncertainty on exporters is not clear with a trimming parameter of 3 percent. By using a trimming parameter of 0.15, we eliminate this period and see the exact relationship in **Figure 7**. As it is clear in the figure, the relationship is far from 10% significance level in the whole sample period for exporter firms. This

*Wald statistics values of the test, H0: financial uncertainty does not Granger-cause MFA score for exporter.* 

uncertainty is stable, there is no significant causality to firm performance.

**172**

**Figure 4.**

*Notes: VAR(2) under SIC, trimming 3%.*

*Wald statistics values of the test, H0: financial uncertainty does not Granger-cause MFA score for nonexporters. Notes: VAR(2) under SIC, trimming 3%.*

#### **Figure 6.**

*Wald Statistics Values of the Test, H0: consumer uncertainty does not Granger-cause MFA score for exporters. Notes: VAR(2) under SIC, trimming 15%.*

might be an expected result since perception of domestic consumers should not have an impact on exporting firms selling abroad. Global financial conditions or global consumer uncertainties are anticipated to have a more impact on exporter firms' performance.

On the other hand, as seen in **Figure 8**, Wald statistics improved extensively after 2009 and exceeded 10% critical levels at some points till 2015 for non-exporter firms. The period between 2010 and 2015 is associated with a sound economic growth with a low-level of interest rates and stable exchange rate. In that period, households' perceptions and future expectations were robust, while their wealth was also improving. This led to soaring consumption expenditures, which increased the sales of non-exporter firms. As a result, the balance sheet performance of nonexporter firms significantly strengthened. We infer from those developments that, as consumers' perception worsens, this does not lead to a causality from consumer

#### **Figure 7.**

*Wald statistics values of the test, H0: consumer uncertainty does not Granger-cause MFA score for exporters. Notes: VAR(2) under SIC, trimming 15%.*

#### **Figure 8.**

*Wald statistics values of the test, H0: consumer uncertainty does not Granger-cause MFA score for nonexporters. Notes: VAR(2) under SIC, trimming 3%.*

uncertainty to the performance of non-exporter firms; nevertheless, as consumer perceptions improve, significant causality from consumer uncertainty index towards MFA scores of non-exporter firms occurs. Besides, for exporter firms, we do not observe a causal relationship in the periods of both improving and worsening consumer perception.

Since the separation of firms depends on their exporting behavior in this analysis, it will be useful to check the effect of trade uncertainty for exporters and non-exporters. Trade uncertainty has increased since 2017 following of Trump's administration and the trade war between the USA and China. In **Figure 9**, trade uncertainty significantly impacts the performance of exporters, with the highest effect seen in 2017–2018. During that period, rising global trade uncertainty adversely influenced the global trade volume. Although high US dollar/Turkish lira exchange rate provided a competitive price advantage for Turkish firms, rising

**175**

**4. Conclusion**

**Figure 10.**

*Notes: VAR(2) under SIC, trimming 3%.*

**Figure 9.**

*VAR(2) under SIC, trimming 3%.*

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength*

global trade uncertainties significantly impacted the balance sheet performance of Turkish exporters. Nevertheless, since global trade was stable in earlier periods, it did not lead to a significant causal relationship. On the other hand, **Figure 10** illustrates that we do not observe such a significant effect on the non-exporters in

*Wald statistics values of the test, H0: trade uncertainty does not Granger-cause MFA score for non-exporters.* 

*Wald statistics values of the test, H0: trade uncertainty does not Granger-cause MFA score for exporters. Notes:* 

In this chapter, we use a time-varying vector auto-regressive model to analyze the relationship between economic uncertainty and firms' balance sheet strength in Turkish economy over the 2005–2019 period. For this purpose, we utilize Multivariate Firm Assessment Score developed by Çolak [36] to calculate Turkish

our model since they do not interact with global trade conditions.

*DOI: http://dx.doi.org/10.5772/intechopen.91860*

*The Relationship between Economic Uncertainty and Firms' Balance Sheet Strength DOI: http://dx.doi.org/10.5772/intechopen.91860*

#### **Figure 9.**

*Banking and Finance*

**Figure 7.**

**Figure 8.**

*Notes: VAR(2) under SIC, trimming 15%.*

**174**

consumer perception.

*exporters. Notes: VAR(2) under SIC, trimming 3%.*

uncertainty to the performance of non-exporter firms; nevertheless, as consumer perceptions improve, significant causality from consumer uncertainty index towards MFA scores of non-exporter firms occurs. Besides, for exporter firms, we do not observe a causal relationship in the periods of both improving and worsening

*Wald statistics values of the test, H0: consumer uncertainty does not Granger-cause MFA score for non-*

*Wald statistics values of the test, H0: consumer uncertainty does not Granger-cause MFA score for exporters.* 

Since the separation of firms depends on their exporting behavior in this analysis, it will be useful to check the effect of trade uncertainty for exporters and non-exporters. Trade uncertainty has increased since 2017 following of Trump's administration and the trade war between the USA and China. In **Figure 9**, trade uncertainty significantly impacts the performance of exporters, with the highest effect seen in 2017–2018. During that period, rising global trade uncertainty adversely influenced the global trade volume. Although high US dollar/Turkish lira exchange rate provided a competitive price advantage for Turkish firms, rising

*Wald statistics values of the test, H0: trade uncertainty does not Granger-cause MFA score for exporters. Notes: VAR(2) under SIC, trimming 3%.*

#### **Figure 10.**

*Wald statistics values of the test, H0: trade uncertainty does not Granger-cause MFA score for non-exporters. Notes: VAR(2) under SIC, trimming 3%.*

global trade uncertainties significantly impacted the balance sheet performance of Turkish exporters. Nevertheless, since global trade was stable in earlier periods, it did not lead to a significant causal relationship. On the other hand, **Figure 10** illustrates that we do not observe such a significant effect on the non-exporters in our model since they do not interact with global trade conditions.

#### **4. Conclusion**

In this chapter, we use a time-varying vector auto-regressive model to analyze the relationship between economic uncertainty and firms' balance sheet strength in Turkish economy over the 2005–2019 period. For this purpose, we utilize Multivariate Firm Assessment Score developed by Çolak [36] to calculate Turkish firms' balance sheet strength. On the other hand, uncertainties in Turkish economy is measured by using four indices: economic uncertainty, financial uncertainty, consumer uncertainty indices which are developed by Cosar and Sahinoz [5], and trade uncertainty index introduced by Ahir et al. [41]. In our empirical analysis, based on Rossi and Wang [42], time-varying Granger causality between uncertainty indices and MFA score is evaluated.

Our first observation is that during heightened volatility periods such as the 2008 global financial crisis and 2018 Turkish domestic FX turmoil, economic uncertainties increased dramatically, and firms' balance sheet performances were negatively affected by those uncertainties. The transmission mechanism was different in those two turbulent periods. The global financial crisis was associated with significant contraction in economic activity, rise in unemployment, and fall in investment and consumption spending in Turkish economy. Those adverse developments in that era caused firms' financial health to worsen. However, during Turkish domestic FX turmoil in August 2018, the sharp depreciation of Turkish lira created a substantial deterioration in the balance sheets of real sector, and many firms were negatively affected since they have problems to repay their FX debts.

Second, while financial uncertainty is found to have a strong effect in exporter firms' performances, we do not observe a significant impact on non-exporter firms' in 2008 and 2018 stress periods. The intuition behind this result is that exporters are expected to be affected more by global financial conditions and heightened exchange rate volatility.

Third, since consumer perceptions about Turkish economy may be more effective on non-exporter (domestic producer) firms, we investigate to role of consumer uncertainty on the performance of Turkish firms. We show that as consumer perceptions improve, significant causality from consumer uncertainty index towards MFA scores of non-exporter firms occurs. However, we do not observe a causal relationship for exporter firms.

Finally, we check the effect of trade uncertainty, which is expected to be an important factor for exporter firms. Our analysis show that the trade uncertainty significantly impacts the performance of exporters, with the highest effect seen in 2017–2018 period during which global trade uncertainty increased dramatically following of Trump's administration and the trade war between the USA and China.
